This and a companion project (Z01 HD001123-12) investigate auditory communication in primates. The overall goal of these studies is to provide a comprehensive understanding of primate auditory communication in terms of development, neural mechanisms, endocrine factors, and social context. Two species, the squirrel monkey and the common marmoset, are the main subjects of study, with additional data collected from other species where appropriate. The present project studies primate communication from a bioacoustic and ethological perspective, focusing on the detailed acoustic structure of vocalizations produced in natural settings, and the relationship of structural differences to differences in age, gender, experience and response selectivity, as well as the broader factors of social context and genetic background. New findings focused on the perceptual side of auditory communication, using playback techniques to examine perceptual preferences as measured by vocally responding to the stimulus. A close-range affiliative vocalization given mainly by adult female squirrel monkeys, the chuck, was tested for responsiveness based on the familiarity of the sound (due to the acoustic details of the chucks of individual vocalizers) to the listener. Chucks recorded from the test group were digitized, then either presented intact, with a frequency-modulated component referred to as a 'flag' removed, or with the flag of another individual (from another social group) electronically spliced onto a chuck instead of the naturally occurring flag for that vocalization. Intact chucks were responded to significantly more than the same chuck with the flag removed, or to chucks with unfamiliar flags spliced onto them. These results suggest that the flag is an important acoustic feature used by females to determine whether they vocally respond to this form of conversational communication. To test whether the acoustic information in flags was sufficient to statistically separate vocalizers, several flag parameters (such as flag duration, peak frequency, etc.) were measured and the values entered into a discriminant analysis. The results indicated that flag parameters alone were sufficient to separate the chucks of vocalizers in a social group.